Stuck processes
Alive is not the same as progressing.
Use progress freshness when a heartbeat alone can miss a frozen worker.
Explore Stuck processesHeartbeatHook solutions
Solutions for Jobs, Services and Private Applications
Start with the operational problem: a missing job, a stuck worker, a private network or a signal that needs a safe return path.
Solution overviewChoose by the problem you need to solve
Solutions pages describe situations in the language engineers use when something is unreliable. They connect that problem to the smallest relevant HeartbeatHook product concept and to a practical guide.
HeartbeatHook does not replace logs, metrics, tracing, backup validation or incident response. The useful design is often layered: liveness evidence, missing-heartbeat policy, progress freshness and application-specific validation each answer a different question.
Keep the failure classes separate
A process can be gone, alive but not checking in, checking in while stuck, or successfully sending a signal that the receiver has not yet processed. The solution pages keep those states distinct so the next implementation decision is clear.
Private applications
Outbound paths can be enough.
Monitor systems behind firewalls or NAT without opening an inbound port.
Explore Private applicationsScheduled work
Execution and completion are different.
Separate scheduler evidence, completion evidence and business validation.
Explore Scheduled workApplication signals
Leave work for a receiver.
Use an authenticated pending mailbox and receiver-controlled polling.
Explore Application signals